Search

Observations with distributed sensors are essential in analyzing a series of human and machine activities (referred to as 'events' in this paper) in complex and extensive real-world environments. This is because the information obtained from a single sensor is often missing or fragmented in such an environment; observations from multiple locations and modalities should be integrated to analyze events comprehensively. However, a learning method has yet to be established to extract joint representations that effectively combine such distributed observations. Therefore, we propose Guided Masked sELf-Distillation modeling (Guided-MELD) for inter-sensor relationship modeling. The basic idea of Guided-MELD is to learn to supplement the information from the masked sensor with information from other sensors needed to detect the event. Guided-MELD is expected to enable the system to effectively distill the fragmented or redundant target event information obtained by the sensors without being overly dependent on any specific sensors. To validate the effectiveness of the proposed method in novel tasks of distributed multimedia sensor event analysis, we recorded two new datasets that fit the problem setting: MM-Store and MM-Office. These datasets consist of human activities in a convenience store and an office, recorded using distributed cameras and microphones. Experimental results on these datasets show that the proposed Guided-MELD improves event tagging and detection performance and outperforms conventional inter-sensor relationship modeling methods. Furthermore, the proposed method performed robustly even when sensors were reduced., Comment: 13page, 7figure, under review

We aim to perform sound event localization and detection (SELD) using wearable equipment for a moving human, such as a pedestrian. Conventional SELD tasks have dealt only with microphone arrays located in static positions. However, self-motion with three rotational and three translational degrees of freedom (6DoF) shall be considered for wearable microphone arrays. A system trained only with a dataset using microphone arrays in a fixed position would be unable to adapt to the fast relative motion of sound events associated with self-motion, resulting in the degradation of SELD performance. To address this, we designed 6DoF SELD Dataset for wearable systems, the first SELD dataset considering the self-motion of microphones. Furthermore, we proposed a multi-modal SELD system that jointly utilizes audio and motion tracking sensor signals. These sensor signals are expected to help the system find useful acoustic cues for SELD on the basis of the current self-motion state. Experimental results on our dataset show that the proposed method effectively improves SELD performance with a mechanism to extract acoustic features conditioned by sensor signals., Comment: ICASSP2024 accepted

The theory of hierarchical equations of motion (HEOM) is one of the standard methods to give exact evaluations of the dynamics as coupled to harmonic oscillator environments. However, the theory is numerically demanding due to its hierarchy, which is the set of auxiliary elements introduced to capture the non-Markovian and non-perturbative effects of environments. When system-bath coupling becomes relatively strong, the required computational resources and precision move beyond the regime that can be currently handled. This article presents a new representation of HEOM theory in which the hierarchy is mapped into a continuous space of a collective bath coordinate and several auxiliary coordinates as the form of the quantum Fokker-Planck equation. This representation gives a rigorous time evolution of the bath coordinate distribution and is more stable and efficient than the original HEOM theory, particularly when there is a strong system-bath coupling. We demonstrate the suitability of this approach to treat vibronic system models coupled to environments., Comment: 18 pages, 5 figures

In the leading order of the large-$N$ approximation, we study the renormalon ambiguity in the gluon (or, more appropriately, photon) condensate in the 2D supersymmetric $\mathbb{C}P^{N-1}$ model on~$\mathbb{R}\times S^1$ with the $\mathbb{Z}_N$ twisted boundary conditions. In our large~$N$ limit, the combination $\Lambda R$, where $\Lambda$ is the dynamical scale and $R$~is the $S^1$ radius, is kept fixed (we set $\Lambda R\ll1$ so that the perturbative expansion with respect to the coupling constant at the mass scale~$1/R$ is meaningful). We extract the perturbative part from the large-$N$ expression of the gluon condensate and obtain the corresponding Borel transform~$B(u)$. For~$\mathbb{R}\times S^1$, we find that the Borel singularity at~$u=2$, which exists in the system on the uncompactified~$\mathbb{R}^2$ and corresponds to twice the minimal bion action, disappears. Instead, an unfamiliar renormalon singularity \emph{emerges\/} at~$u=3/2$ for the compactified space~$\mathbb{R}\times S^1$. The semi-classical interpretation of this peculiar singularity is not clear because $u=3/2$ is not dividable by the minimal bion action. It appears that our observation for the system on~$\mathbb{R}\times S^1$ prompts reconsideration on the semi-classical bion picture of the infrared renormalon., Comment: 26 pages, 1 figure, the final version to appear in PTEP

Due to its exceptional lithium storage capacity silicon is considered as a promising candidate for anode material in lithium-ion batteries (LIBs). In the present work we demonstrate that methods of the soft X-ray emission spectroscopy (SXES) can be used as a powerful tool for the comprehensive analysis of the electronic and structural properties of lithium silicides Li$_{x}$Si forming in LIB's anode upon Si lithiation. On the basis of density functional theory (DFT) and molecular dynamics (MD) simulations it is shown that coordination of Si atoms in Li$_{x}$Si decreases with increase in Li concentration both for the crystalline and amorphous phases. In amorphous a-Li$_{x}$Si alloys Si tends to cluster forming Si-Si covalent bonds even at the high lithium concentration. It is demonstrated that the Si-L$_{2,3}$ emission bands of the crystalline and amorphous Li$_{x}$Si alloys show different spectral dependencies reflecting the process of disintegration of Si-Si network into Si clusters and chains of the different sizes upon Si lithiation. The Si-L$_{2,3}$ emission band of Li$_{x}$Si alloys become narrower and shifts towards higher energies with an increase in Li concentration. The shape of the emission band depends on the relative contribution of the X-ray radiation from the Si atoms having different coordination. This feature of the Si-L$_{2,3}$ spectra of Li$_{x}$Si alloys can be used for the detailed analysis of the Si lithiation process and LIB's anode structure identification.